Aero upgrades ranked by watts per euro

At walking pace almost all the work you do goes into rolling resistance and gravity. By the time you are moving at a normal club-run speed, air has taken over completely. On the flat at 35 km/h, a rider with an unremarkable position and decent tyres is spending roughly four fifths of their power pushing air out of the way and the rest on everything else combined.

That is the reason the aero industry exists, and it is also the reason most of what it sells you is in the wrong order. The single largest object in your frontal area is you. Every component you can buy is trying to improve the small remainder.

This post ranks the things you can change by what they cost per watt saved, using one consistent set of assumptions so the comparisons are fair. The numbers are worked examples, not promises.

The cubic relationship, in actual watts

Aerodynamic drag force rises with the square of your speed. The power needed to overcome it rises with the cube, because power is force times velocity. That is why the last few km/h are so expensive and why the gains from a drag reduction get bigger the faster you already are.

Here is the standard road cycling power model — the one validated against field measurements by Martin and colleagues — run for an 80 kg rider-plus-bike with a drag area of 0.32 m² and a rolling resistance coefficient of 0.005.

Read the gap between the two series. Going from 25 to 35 km/h costs about 129 W, and almost all of it is air. This also means the same percentage saving in drag is worth more watts the faster you ride — which is why aero kit matters far more in a time trial than on a 20 km/h gravel ramble.

Where the drag actually is

Wind-tunnel and CFD studies that partition drag between the rider and the machine keep landing in the same place: the rider is most of it. Reviews of competition cycling aerodynamics put the body at roughly 70–80% of total drag, with the bicycle making up the rest. The individual body segments have been mapped too — the legs, being large, moving and badly shaped for flow, dominate, with the head, arms and torso next.

Once you have seen that bar, the shopping order writes itself. Anything that changes the shape of the 72% beats anything that changes the shape of the 11%.

The ranking

Each row below takes the 180 W our example rider spends on air at 35 km/h and applies a plausible percentage reduction in drag area, then divides the price by the watts. Percentage ranges are deliberately wide, because the honest answer for most of them is "it depends on where you are starting from".

The spread between the top and the bottom of that list is roughly two orders of magnitude. Position is somewhere between free and the price of a bike fit. A frame is a few thousand euros for a handful of watts. Both are real; only one of them is a sensible first purchase. If you have never had your contact points measured properly, bike fit basics is the place to start, because a fit that lets you hold a lower position for three hours is worth more than a fit that merely makes you low for ten minutes.

Why position beats everything, and why it is still hard

CFD and full-scale wind-tunnel comparisons of the same rider in an upright position, a dropped position and a time-trial position show drag area falling substantially at each step — the difference between sitting up on the tops and getting properly low is comfortably larger than the difference between any two wheelsets you could buy.

The catch is that the tunnel measures a static pose and your ride does not. Position gains only count for the fraction of the ride you actually hold them. Riders routinely buy 15 watts in a tunnel and then spend 80% of the ride sat up, because their hamstrings, their lower back or their neck will not take it. Lower back trouble is the most common failure mode here, and it is worth reading what actually causes lower back pain on the bike before you slam the stem.

The fix is boring: change position in small increments, and spend time in the new one when the consequences of being uncomfortable are low. A trainer is ideal for this, because you can hold a position without traffic, junctions or descents pulling you out of it — Moveee Indoor runs a full session in a browser tab, so you can do your Tuesday endurance ride in the new position and find out on the third week whether your back has stopped complaining. It is free while in alpha.

Clothing: the best value item you can actually buy

Fabric research is one of the few parts of cycling aerodynamics with a solid open literature. Wind-tunnel testing of jersey fabrics at cycling speeds shows that surface roughness and knit structure change drag measurably on cylinders sized like human limbs, and that the optimal roughness depends on the local flow speed — which is why sleeve fabric on a fast rider is often different from torso fabric.

What this means practically is duller than the marketing. A close-fitting jersey with sleeves that do not flap is most of the available gain. The step from that to a purpose-built suit is smaller, and the published evidence for the size of that step is thin outside commercial testing. Buy the fit first.

Helmets and wheels: real, small, and conditional

Wind-averaged testing of time-trial helmets found meaningful differences between models, but also that ranking depends on yaw angle and head position — a helmet with a long tail is fast when your head is where the designer assumed and slow when you drop your chin to look at your stem. CFD work on time-trial helmets reaches the same conclusion from the other direction.

Wheels are the classic example of a conditional gain. CFD and wind-tunnel studies of spoked and deep-section wheels show that the interesting behaviour happens at yaw, where a deep aerofoil rim can generate a side force with a small forward component. At zero yaw, on a still day, the difference between a 30 mm and a 60 mm front rim is a few watts. On a gusty day at 10–15 degrees of yaw it is larger — and the handling is worse. If you want to know how often you will actually get that crosswind, reading the wind forecast is more use than another tunnel graph.

How fast the returns die

There is an arithmetic reason this list flattens out so quickly. Speed at fixed power scales roughly with the cube root of the drag reduction. Cut your drag area by 10% and you go about 3.5% faster. Cut it by another 10% and you gain roughly the same fraction again — but you have now spent far more to get it, because you took the cheap 10% first.

Work it through for a 40 km ride at 35 km/h, which takes 68.6 minutes. A 10% drag reduction brings that down to about 66.3 minutes: a bit over two minutes. A further 5% saves roughly another 50 seconds. The third 5% saves less than that again, and by then you are into four-figure component prices.

What to do with this

• Spend nothing first. Get low, get your elbows bent, get your hands on the hoods rather than the tops, and stop the shoulders rolling side to side. Then check you can still hold it at hour three.
• Then buy clothing that fits. It is the cheapest thing on the list with a real published basis behind it.
• Then a helmet, if yours is old and vented like a colander. Modest, but cheap per watt.
• Only then wheels. Front wheel first if you are buying one at a time.
• Frames last. Buy one because you want a new bike, not because of a drag figure.
• Do not forget the boring stuff. Rolling resistance is the second-biggest term in the equation and it is much cheaper to improve than drag — tyre pressure and a clean, waxed chain together can be worth as much as a wheelset, for the price of a coffee.

The uncomfortable summary is that the fast riders you see are not fast because of their kit. They are low, they are lean, they are comfortable in a position that would hurt you, and their tyres are set up properly. All of that is available before you open your wallet.